Ice Age research. Milankovitch cycles, Milankovitch curves, Milankovitch insolation, Milankovitch theory, Milankovitch hypothesis.?

Transcription

1 Ice Age research Rev. Prof. Buckland Milankovitch cycles, Milankovitch curves, Milankovitch insolation, Milankovitch theory, Milankovitch hypothesis.? Milutin Milankovitch

2 Milutin Milankovitch ( ) I do not consider it my duty to give an elementary education to the ignorant, and I have also never tried to force others to apply my theory, with which no one could find fault. Such an assertion can only come from someone with amazing courage, since he dares to write of things about which he knows nothing. (Milankovitch, 1941)

3 Orbital parameters Earth s orbit varies over time due to influence of the Sun, Jupiter, and the Moon. Eccentricity (ellipticity) ~100 kyr, 400 kyr Obliquity (tilt) ~41 kyr Precession (wobbly top) ~19, 23 kyr

4 Note also Kepler s law s 1. The orbit of every planet is an ellipse with the Sun at a focus. 2. A line joining a planet and the Sun sweeps out equal areas during equal intervals of time. 3. The square of the orbital period of a planet is directly proportional to the cube of the semi-major axis of its orbit. zip round when close, slow down when far away Important for seasonallyintegrated aspects of climate Johannes Kepler, Are all a consequence of Newtonian gravitation Principia 1687 (also calculus, laws of motion, tides, planetary masses, comets, precession of equinoxes, etc.)

5 Insolation variations at 65N - Summer Solstice Insolation (Wm -2 ) Time (kyr)

6

7 δ18o ( Integrated summer insolation (GJ per m2) 5 21 Dec. 65 S insolation (W per m2) c 21 June 65 N insolation (W per m2) b Obliquity (degrees) Peak summer vs. Integrated summer insolation , ,000 1,000 1,200 2,2001,200 2,4001,400 2,60 1, , ,400 2, ,600 2, ,800 2, ,000 1,000 3,000 Age (thousands of years ago) ds of years ago) Age (thou & Huybers, From 3Raymo interval. to 1 million years ago, lity. A 3-million-year record of O varies primarily at of the 41 interval. From 3 and to 1 solar million years ago,aδ3-million-year Ice-age climate variability. record mpared with integrated summer year period characteristic of obliquity a year characteristic of obliquity andwith integrated insolation. Fro. 8) (aperiod ); orbital obliquity (blue) compared integrated summer 1 million years ago to the present, long onwhich for the Northern Hemisphere more important for ablation? 1 million to the present, longer cycles of climate change, w ); andago summer insolation for the Northern Hemisphere on (red)3 (isbyears

8 Milutin Milankovitch ( ) (Milankovitch, 1941) Glaciations correspond to summer insolation minima

9 What does an ice age look like? Reconstruction of land and sea ice 21,000 years ago February July CLIMAP reconstruction

10 What does an ice age look like? (Eastern Greenland -Steve Porter)

11 Western Greenland

12 What does an ice age look like? Puget sound lobe Ralph Haugerud Harvey Greenberg

13 Harvey Greenberg

14 Harvey Greenberg

15 Harvey Greenberg

16 Harvey Greenberg

17 Inferring global ice volume in the past 16 O isotopes 18 O isotopes ice ocean sediment land fractionation of isotopes during evaporation and precipitation favours light isotopes being transported to high latitudes. ocean becomes enriched in heavy isotopes. signal can be measured in ocean sediments.

18 Inferring global ice volume in the past proxy for ice volume over the last 2.5 million years more ice less ice 100 kyr ice ages The Pleistocene (ish)

19 The SPECMAP ice volume time series Composite stack of many δ 18 O deep sea cores (~20) Imbrie et al., 1984

20 The SPECMAP ice volume time series and June insolation at 65N (upside down) maximum correlation of -0.4 with a 6 kyr lag of ice volume behind insolation more ~100 kyr variability in ice volume than in insolation

21 What people say about this 6 kyr lag is due to dynamical response of ice sheets CO 2 leads ice volume by ~6 kyr Tropical temperature lead ice volume by ~6 kyr => CO 2 /SSTs force climate change S.H. temperatures lead ice volume by ~6 kyr => S.H. source of deglaciation mechanism It takes 6 kyr for climate signal to reach the N.H. => role of deep ocean/chemistry BUT, ice volume is a bad climate variable

22 Rate of change of ice volume d(volume)/dt more directly related to high latitude insolation

23 Rate of change of ice volume d(volume)/dt more directly related to high latitude insolation maximum correlation of -0.8 at zero lag

24 Rate of change of ice volume d(volume)/dt more directly related to high latitude insolation major difference is large negative rates of change during major deglaciations Terminations coincide with insolation maxima - points to insolation trigger

25 The role of CO 2 : Changes in CO 2 lead changes in ice volume

26 The role of CO 2 : The cross spectrum between CO2 and dv/dt CO 2 lags CO 2 leads Variations in melting precede variations in CO 2!

27 Summary Waxing and waning of global ice volume strongly controlled by high latitude, northern hemisphere summertime insolation (i.e., Milankovitch s original idea, sort of) CO2, tropical SSTs lag dv/dt - not the primary driver of ice-sheet variations Insolation trigger for collapse? Reason for deglaciation still unknown Changes the question from does orbital forcing affect global ice volume (it does), to what causes the big deglaciations?

28 Other evidence of Milankovitch ice age cycles Isotope records in stalagmites in Chinese caves (recording O2 isotopes in precipitation)

29 Other evidence of Milankovitch ice age cycles Bathymetry of spreading ridges (but controversial) Crowley et al., 2015 Blue: actual Green: filtered Black: model

30 How an ice sheet works (roughly): 1. Force balance on a volume within ice sheet Surface slope leads to pressure gradient P x P x+dx ice land σ ij ocean Pressure gradient balances shear stress: σ ij = z s z p dz x

31 How an ice sheet works (roughly): 2. Ice responds to stress by deforming (creep flow) P x P x+dx ice land σ ij ocean Glen s flow law: relates strain rate to applied stress ε (~ du dz ) = A(T) σ 3 ij Strain rate (Stress) 3

32 ice flow is basically a very nonlinear diffusion equation How an ice sheet works (roughly): 3. Equilibrium state is a flux balance ablation zone F(x) Accumulation zone ice land ocean Steady state mass balance: $ Flux of ice H 5 dh ' s & ) % dx ( 3 x 0 = (snowfall melting) dx

33 Accumulation Greenland - average accumulation ~30 cm/year. Antarctica - average accumulation ~10 cm/year. Clausius-Clapeyron relationship: moisture content decreases exponentially with height. e-folding scale height at high latitudes ~ 2 km. most of Antarctica, Greenland are above this height. As ice sheet grows, average snowfall decreases strongly strong negative feedback as ice sheet grows

34 Ohmura et al., 1996 Greenland Ablation vs. Summer temperature ~0.5 m/yr per o C of T JJA

35 Ablation Ablation rate summer temperature. T Rule of thumb 1 o C JJA 1myr -1 melting. Ice sheet margin has parabolic shape. Total ablation T 3. T+ΔT Ice sheets are very sensitive to summertime temperatures!

36 Ice sheet sensitivity: accumulation rate vs. ablation rate Contours of ice sheet length (in km) need a big accumulation rate change to offset a small temperature change.

37 Dynamic ice sheet model evolution (Hyo-Seok Park) Ice sheet forced with 40 kyr sinusoidal forcing +/- 5 o C

38 Ice volume response for different forcing periods 20 kyr forcing ice volume 40 kyr 100 kyr

39 Histograms of the rate of change of ice volume

40 Ice sheet: white noise only

41 Ice sheet: sinusoid plus noise

42 Use insolation variations to force ice sheet model Forcing = response

43 Comparison of model dv/dt with model forcing maximum correlation 0.4 at zero lag

44 Use insolation variations to force ice sheet model plus noise Forcing (plus white noise) = response

45 Comparison of model dv/dt with model forcing plus noise maximum correlation 0.35 at zero lag

46 Conclusions Models and observations agree that rate of change of ice volume is directly related to insolation variations. (up to 60% of variance) The Milankovitch hypothesis (formulated properly) does, in fact, explain a lot about climate change, but noise is an important part too. Ice sheets are very nonlinear beasts. How do large ice sheets collapse?

47

48 List of extras Rectification of precession forcing. Spectra of models and obs. Cross spectra of dv/dt vs insolation. Equations comparing different theories. CO 2 and Vostok. Ohmura figure of glacial ablation. Haug et al results

49 Rectification of 19 and 23 k forcing Forcing Response

50 Rectification of 19 and 23 k forcing: spectral estimates Forcing 1/100 1/23 1/19 (plus noise) Response

51 Rectification of 19 and 23 k forcing: spectral estimates Forcing 1/100 1/23 1/19 (plus noise) Response

52 Comparison of data and model spectra 1/100 1/41 1/23 1/19

53 Comparison of data and model spectra 1/100 1/41 1/23 1/19 1/100 1/41 1/23 1/19

54 Ice volume, length, and height for 40 kyr forcing

55 Cross-spectrum SPECMAP dvol/dt with June 65N

56 Cross-spectrum Huybers dvol/dt with June 65N broad window

57 Cross-spectrum Huybers dvol/dt with June 65N narrow window

58 Ice volume evolution equation: τ >> 2π / f d dt volume(t) + volume(t) τ = insolation(t) vs. volume(t) = insolation(t 7kyr)

59 Haug et al., 2005: Pacific moisture and glacial inception?

60

61 So why is this dubious? Ice volume records have been tuned to orbital parameters! Age records are created by maximizing co-variance in 20 kyr, 40 kyr bands, assuming fixed (& different) phases with obliquity and precession. So argument linking dv/dt to insolation is nearly circular (but not quite)

62 Ice volume record independent of orbital tuning (Peter Huybers, MIT) 26 ocean cores anchor each core at LGM and last magnetic reversal: => gives separate time scale for each core fix timing of each deglaciation at the average value for all cores, and re-tweak the ages. Ice volume record independent of orbital assumptions

63 Ice volume record independent of orbital tuning (Peter Huybers, MIT)

64 Rate of change of ice volume record and insolation upside down maximum correlation of -0.4 at zero lag

65 Rate of change of ice volume record and insolation upside down maximum correlation of -0.4 at zero lag deglaciations stand out more than SPECMAP

66 Other ways of getting the same answer: Insolation at any latitude and time can be described by a linear combination of obliquity and precession i.e., S(x,t) = a 1. precession(t-τ 1 ) + a 2. oblquity(t - τ 2 ) Bad news: Cannot unambiguously attribute signal to forcing at a particular latitude and time of year. Good news: Can ask what combination of obliquity and precession is the best explanation of ice volume Answer: Same as June 65N (or something close)

67 Linear regression of orbital parameters onto d(vol)/dt

68 Linear regression of orbital parameters onto d(vol)/dt

69 Linear regression of orbital parameters onto d(vol)/dt

70 Linear regression of orbital parameters onto d(vol)/dt